SEARCH

SEARCH BY CITATION

Keywords:

  • penile neoplasm;
  • squamous cell carcinoma;
  • competing risk;
  • survival;
  • Surveillance, Epidemiology, and End results (SEER) program

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

What's known on the subject? and What does the study add?

The European Association of Urology (EAU) guidelines recommend an inguinal lymph node dissection (ILND) in patients with T1G2-3 squamous cell carcinoma of the penis (SCCP). To date, only four series reported the rates of cancer-specific mortality (CSM) after primary tumor excision (PTE) without an ILND in patients with T1 clinically node-negative (cN0) SCCP.

We examined CSM rates in cN0 patients with T1G1-3 SCCP, in whom an ILND was not performed, relying on competing-risks analyses.

OBJECTIVE

  • • 
    To quantify and compare cancer-specific mortality (CSM) and other-cause mortality (OCM) in individuals with stage T1G1–3 clinically node-negative (cN0) squamous cell carcinoma of the penis (SCCP) since there is no consensus regarding the need for an inguinal lymph node dissection (ILND) in patients with T1G2–3 cN0 SCCP.

METHODS

  • • 
    Relying on the Surveillance, Epidemiology and End Results database, we identified 655 patients diagnosed with primary SCCP between 1988 and 2006.
  • • 
    Cumulative incidence plots were used to graphically depict the effect of CSM relative to OCM.
  • • 
    Competing-risks regression analyses were used to quantify the risk of CSM or OCM after adjusting for age, race, tumour grade and surgery type.

RESULTS

  • • 
    The 5-year CSM rates after a primary tumour excision without ILND were 2.6%, 10.0% and 15.9% in patients with respectively T1G1, T1G2 and T1G3 cN0 SCCP.
  • • 
    The 5-year OCM rates were 29.5%, 27.3% and 29.3% in patients with respectively T1G1, T1G2 and T1G3.
  • • 
    Age failed to provide additional stratification.

CONCLUSIONS

  • • 
    The CSM rate was highest in T1G3 patients and appears to justify ILND.
  • • 
    Conversely, the CSM rate was lowest in T1G1 patients, which justifies active surveillance in this patient subset.
  • • 
    A moderate CSM rate at 5 years was recorded for T1G2 patients, which brings into question the benefits of ILND.

Abbreviations
SCCP

squamous cell carcinoma of the penis

EAU

European Association of Urology

ILND

inguinal lymph node dissection

CSM

cancer-specific mortality

PTE

primary tumour excision

OCM

other-cause mortality

SEER

Surveillance, Epidemiology, and End Results.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Squamous cell carcinoma of the penis (SCCP) is a rare disease with an incidence of 0.1 to 0.9 new cases per 100 000 males per year [1]. The European Association of Urology (EAU) guidelines recommend an inguinal lymph node dissection (ILND) in patients with T1G2–3 tumours [2]. Conversely, in North America an ILND is more often used only in T1G3 patients [3]. These recommendations depend on several other patient characteristics, such as age and comorbidities, which can influence the clinical decision to perform or not to perform an ILND. It is conceivable that only a proportion of patients with T1G2–3 tumours undergo an ILND. Unfortunately, exact numbers are not available. However, what is known is that only between 39.1% and 49.0% of patients with SCCP stages T1–4 undergo an ILND [4–7].

Lack of adherence to ILND recommendation may stem from the relatively unknown natural history of SCCP after removal of the primary tumour. Lack of familiarity with the natural history may be equalled by some clinicians with indolent natural history and may lead to an underestimation of the cancer-specific mortality (CSM) risk. To date, only four series reported the rates of CSM after primary tumour excision (PTE) without an ILND in patients with T1 clinically node-negative (cN0) SCCP (Table 1) [8–11]. In those studies, the number of patients at risk ranged from eight to 24 and the rates of CSM ranged from 0.0% to 17.6%. The error margins associated with these CSM estimates are wide and the CSM rates may be of limited reliability. To address this limitation, we decided to examine CSM rates in cN0 patients with T1G1–3 SCCP in whom an ILND was not performed. In distinction to previous studies, our cohort represents a population-based tumour registry. Moreover, our methods rely on competing-risks analyses, which allow us to quantify the effect of other-cause mortality (OCM) in the light of CSM. Our main hypothesis stated that an important proportion of patients with T1G2–3 SCCP might succumb to SCCP.

Table 1. CSM in patients with cN0 groin T1 SCCP treated with excisional biopsy, partial or total penectomy without initial lymphadenectomy
 Tumour gradeNo. of patientsTime point (year)CSM (%)
Fraley et al. [8]G1850
McDougal [9]G1–21570
Theodorescu et al. [10]G1–3244.412.5
Theodorescu et al. [10]G184.412.5
Schlenker et al. [11]G2176.517.6

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

STUDY POPULATION

The Surveillance, Epidemiology, and End Results (SEER) database of the National Cancer Institute programme covers approximately 26% of the US population and is considered representative of the USA regarding demographic composition as well as cancer incidence and mortality [12]. The reliability of the SEER database is high and is continuously improved [13,14]. The 17 SEER registries include Alaska, Metropolitan Atlanta, Greater California, Los Angeles, San Francisco – Oakland, San Jose – Monterey, as well as Connecticut, New Jersey, Detroit (Metropolitan), Iowa, Kentucky, Utah, Louisiana, New Mexico, Rural Georgia, Seattle (Puget Sound) and Hawaii. Within the 17 SEER tumour registries, we identified all men treated for primary SCCP between 1988 and 2006 who underwent PTE (excisional biopsy, partial or total penectomy) according to two diagnostic codes: the tenth revision of the International Classification of Disease for Oncology second edition (ICD-O-2) (C60.0–60.9) and the ICD-O-3 codes for histological subtype (squamous cell carcinoma type, ICD-O-3: 8070–8076).

Only patients with squamous cell histology and cN0 groins were included. Tumour grade was stratified according to the SEER database between grades 1, 2 and 3. We then relied on the SEER primary tumour stage and grade to stratify patients between T1G1, T1G2 and T1G3 stage/grade groupings [15]. The 2002 version of the TNM staging manual was used instead of the 2009 version due to lack of information allowing us to convert the SEER data to the 2009 version in T1 substages [16]. Exclusions consisted of unknown, T2 or higher tumour stage, unknown tumour grade, presence of clinically involved inguinal lymph nodes and/or the presence of distant metastases. Similarly, all patients subjected to ILND at the time of PTE were also excluded.

PTE was considered as the start time of observation. For the purpose of this analysis, deaths from penile cancer were coded as cancer-specific events according to the SEER-specific cause of death recode (28030). All other deaths were considered as OCM.

STATISTICAL ANALYSES

The t test and chi-squared test were used to compare the statistical significance of differences in respectively means and proportions. A substantial proportion of patients with SCCP die as a result of competing causes of mortality. To account for competing mortality, we used competing-risks regression analysis, as described by Fine and Gray [17]. This accounts for the effect of OCM and provides the most unbiased estimates of CSM.

Cumulative incidence plots were used to graphically depict the effect of CSM after accounting for OCM in the overall population. Additionally, age- and grade-stratified cumulative incidence plots examined the effect of CSM in patients with T1 cN0 SCCP. The Gray test [18] was used for comparisons of CSM rates between different groups.

Finally multivariable competing-risks regression models were used to test the effect of grade on CSM rates. Covariates consisted of age (≤69 vs ≥70 years), race (white vs black vs other) and treatment type (excisional biopsy vs partial penectomy vs total penectomy).

In order to provide more stable estimates of survival curves, we relied on the Epanechnikov kernel smoother [19] incorporated in the R statistical package system (R Foundation for Statistical Computing, Vienna, Austria). All other statistical analyses were performed using the Statistical Package for Social Science, version 15.0 (SPSS®, Chicago, IL, USA), statistical software and S-PLUS Professional, version 1 (Mathsoft, Seattle, WA, USA). All tests were two-sided, with a significance level set at 0.05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The study population consisted of 655 cN0 patients who underwent PTE for T1 SCCP between 1988 and 2006 (Table 2). The average age was 67.0 years (median 69.0). White race predominated (86.4%). Most had tumour grade 2 (43.5%) or 1 (42.0%). Overall, 32.4%, 59.8% and 7.8% of patients had respectively an excisional biopsy, a partial penectomy or a total penectomy. Mean and median follow-up were 47.1 and 34.0 months. Overall, 6.4% of patients died of penile cancer and 30.5% died of competing causes. Of patients who died of other causes, the three most common causes were heart disease (8.4%), chronic obstructive pulmonary disease (2.7%) and cerebrovascular disease (2.4%).

Table 2. Descriptive characteristics of the study population treated for T1 SCCP with PTE (1988–2006)
VariablesOverall population
Number of patients655 (100%)
Age (years)
 Mean (median)67.0 (69.0)
 Range28–102
Race
 White566 (86.4)
 Black47 (7.2)
 Other42 (6.4)
Tumour grade
 Grade 1275 (42.0)
 Grade 2285 (43.5)
 Grade 395 (14.5)
Treatment type
 Excisional biopsy212 (32.4)
 Partial penectomy392 (59.8)
 Total penectomy51 (7.8)
Follow-up (months)
 Mean (median)47.1 (34.0)
 Range0.1–209
Vital status at last contact
 Alive413 (63.1)
 Deceased due to penile cancer42 (6.4)
 Deceased due to other causes200 (30.5)

Cumulative incidence plots (Fig. 1) illustrate CSM and OCM rates for the entire population (n= 655). Within the study cohort, at 5 years after PTE, CSM and OCM rates were respectively 7.6% and 28.6%. Figure 2 shows the CSM and OCM rates after stratification according to tumour grade. The 5-year CSM rates for tumour grades 1, 2 and 3 were respectively 2.6%, 10.0% and 15.9% (P< 0.005). The 5-year OCM rates for tumour grades 1, 2 and 3 were respectively 29.5%, 27.3% and 29.3% (P> 0.2).

image

Figure 1. Cumulative incidence plots depicting CSM and OCM in patients with cN0 inguinal lymph nodes and stage T1 SCCP treated with PTE without lymphadenectomy (n= 655). Black indicates CSM; grey indicates OCM.

Download figure to PowerPoint

image

Figure 2. Cumulative incidence plots depicting CSM and OCM in patients with cN0 inguinal lymph nodes and stage T1 SCCP treated with PTE without lymphadenectomy after stratification according to tumour grade: (A) grade 1; (B) grade 2; (C) grade 3. Black indicates CSM; grey indicates OCM.

Download figure to PowerPoint

In univariable competing-risks regression analyses, tumour grade achieved statistical significance (Table 3). Specifically, patients with G2 and G3 SCCP were respectively at 3.3-fold (95% CI 1.4–7.7, P= 0.005) and 5.8-fold (95% CI 2.3–14.8, P< 0.001) higher risk of CSM than their G1 counterparts. Patients treated with total penectomy were at 3.4-fold higher risk (95% CI 1.4–8.6, P= 0.008) of CSM than their counterparts treated with excisional biopsy. Neither age stratified according to the median (P= 0.3) nor race (white vs black vs other) (P= 0.1) achieved statistical significance.

Table 3. Univariate and multivariate competing-risks regression models addressing CSM after accounting for OCM, in the overall population of patients treated for cN0 T1 SCCP with a PTE without initial lymphadenectomy
 Univariate analysisMultivariate analysis
HR (95% CI) P HR (95% CI) P
  1. HR, hazard ratio.

Tumour grade
 Grade 2 vs Grade 13.3 (1.4–7.7)0.0053.3 (1.5–7.5)0.004
 Grade 3 vs Grade 15.8 (2.3–14.8)<0.0016.3 (2.5–16.0)<0.001
Age
 ≥70 vs ≤690.7 (0.4–1.4)0.30.7 (0.4–1.4)0.3
Race
 Black vs White5.6 (0.7–48.5)0.18.1 (0.97–67.7)0.054
 Other vs White2.9 (0.4–21.1)0.33.6 (0.5–26.3)0.2
Surgery type
 Partial penectomy vs excisional biopsy1.1 (0.6–2.3)0.71.1 (0.5–2.2)0.8
 Total penectomy vs excisional biopsy3.4 (1.4–8.6)0.0083.3 (1.3–8.3)0.01

In multivariable competing-risks regression analysis, tumour grade achieved independent predictor status (Table 3). Specifically, patients with G2 and G3 SCCP were respectively at 3.3-fold (95% CI 1.5–7.5, P= 0.004) and 6.3-fold (95% CI 2.5–16.0, P< 0.001) higher risk of CSM than their G1 counterparts. Patients treated with total penectomy were at 3.3-fold higher risk (95% CI 1.3–8.3, P= 0.01) of CSM than their counterparts treated with excisional biopsy. Age (P= 0.3) and race (P= 0.054) failed to reach independent predictor status.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Our results demonstrate 10.0% and 15.9% CSM rates at 5 years after PTE without an ILND in patients with T1G2 and T1G3 cN0 SCCP respectively. Conversely, the 5-year CSM rate in patients with T1G1 cN0 SCCP was 2.6%.

The 15.9% CSM rate recorded in T1G3 patients indicates that one individual in six will eventually succumb to SCCP. This observation will probably convince most patients and clinicians to consider ILND in this subset. This rate supports the frequent use of ILND in North American patients with T1G3 SCCP. Similarly, the T1G3 CSM rate supports the EAU recommendation for an ILND in this patient group [2]. The consideration of ILND is based on the assumption of its therapeutic benefit in the presence of SCCP with clinically indolent groins, despite the presence of established lymph node metastases. Unfortunately, the magnitude of the risk reduction related to the use of ILND is not known.

The recorded CSM rate (10.0%) for T1G2 patients indicates that one individual in 10 will eventually succumb to SCCP. The EAU SCCP treatment guidelines again recommend an ILND in this patient group [2]. There are no North American recommendations or large-scale data to either support or refute the need for an ILND. Our findings indicate that 27.3% of T1G2 patients will succumb to other causes at 5 years after PTE when an ILND is not performed, provided that no CSM event occurs. Our univariable and multivariable analysis failed to identify age as an important stratifying variable that could help with clinical decision making regarding the need for ILND. The non-randomized nature of the SEER database prevents us from being able to provide a relative benefit of an ILND. Therefore, it is up to the clinician and patient to decide whether the 10.0% risk of CSM without an ILND may be reduced if an ILND were performed and to what extent this decrease would reduce the CSM risk. Unfortunately, non-randomized controlled trials have never addressed this issue and none is under way. In consequence the CSM and OCM rates for T1G2 patients corroborate the controversy regarding the benefit of ILND in the context of 10.0% CSM and 27.3% OCM risk at 5 years after PTE.

The 2.6% CSM rate recorded in T1G1 patients suggests a relatively low risk mortality from SCCP. Specifically, only one out of 38 individuals will succumb to SCCP when an ILND is not performed after PTE. Conversely, the OCM rate of 29.3% indicates that one in four individuals will succumb to OCM. This comparison will probably convince most patients and clinicians to consider active surveillance in this patient cohort.

Although our data and plots provide a valuable perspective on SCCP CSM rates, they need to be interpreted with some caution. For example, not performing an ILND may not prevent the spread of metastasis to distant sites, especially if the ILND is performed in the context of lymph node metastasis. Our results are not compared with a series that did undergo an ILND. However, it may be stated that the SEER database records the highest stage for each patient. In consequence, for patients subjected to an ILND, the cN stage is not available as the pathological N stage takes precedence. In consequence, a comparison between cN0 patients treated with ILND vs no ILND cannot be attempted using the SEER database. Additionally, our database does not contain clear indications for omitting an ILND. It can be assumed that, in most instances, the reason was the absence of suspicious clinical or radiographic findings. Moreover, age and comorbidities may also have affected the decision. Finally, the consideration of morbidities and complications associated with an ILND may have further affected the decision. Unfortunately the rationale for omitting an ILND could not be accounted for in the current study nor was it available in previous studies [8–11]. To date, no randomized ILND trial has been performed. In consequence, the reported rates represent the best available estimation of CSM rates for patients with T1G1–3 cN0 SCCP in whom an ILND was not performed. Nonetheless, they need to be interpreted in the light of the enumerated limitations.

In the current study, 42 patients (97.6%) died of penile cancer. Of these, 41 died in the first 5 years (Table 2). This rate is in agreement with EAU guidelines on the follow-up of patients with SCCP [2]. The guidelines suggest clinical follow-up up to 5 years only. After that period, only regular self-examination with clinical follow-up as needed is recommended.

In a cN0 population, a modified ILND as described by Catalona in 1988 is associated with less morbidity than a radical ILND [20]. The morbidities of a modified ILND consist of postoperative skin necrosis (2.5%–4.5%), wound infections (0.8%–14.2%), oedema (3.0%–20%) and lymphocele (0%–30.0%) [21–24]. A modified ILND may represent a worthwhile trade-off in some T1G2 or even T1G3 patients. A modified ILND can be extended to a radical ILND if positive nodes are found. The known complications of ILND are generally manageable. Nonetheless, they are considerable for those who experience them and lead to non-negligible morbidity and quality-of-life detriments. To obviate potential complications, other assessment modalities such as dynamic sentinel lymph node biopsy have emerged. This technique represents an alternative to a modified ILND. The rate of a false negative finding relative to a radical ILND was 6.9% [25]. Associated complication rate was 4.7%. Therefore, this alternative to ILND does warrant consideration in pT1G2 and even some T1G3 patients to whom the idea of a modified or radical ILND is not acceptable [26]. Similar to the modified ILND, the dynamic sentinel lymph node biopsy can be extended to a radical ILND in the presence of positive lymph nodes. The complexity of the treatment considerations regarding the need for, and the extent of, an ILND indicates that patients with SCCP should ideally be managed at tertiary care centres.

Our study is not devoid of limitations. First, lack of central pathology review might represent one weakness. Central pathology review could have contributed to higher accuracy of pathologically assessed variables [27]. Second, the retrospective nature of the study represents a limitation that is shared with all previous studies addressing SCCP outcomes [8–11]. Third, total penectomy for a T1 SCCP is very unusual, but was the case for 7.8% of patients in our study. Some of them may have been misclassified because of the reliability of the SEER database which is high but not perfect. One other explanation is that SCCP is a rare pathology and perhaps surgeons have done a total penectomy in ignorance of the guidelines. Fourth, we do not know for certain that patients did not have other non-invasive or minimally invasive procedures such as a sentinel node procedure. This is another shortcoming of the SEER database, even if its reliability is high. Fifth, other variables, such as invasion of perineural spaces by tumour or lymphatic venous embolization, may also be considered as predictors of CSM [28,29]. However, these variables were not available in the SEER database. This represents a rate-limiting factor and prevented us from examining potentially important variables other than that of tumour grade. Moreover, in the event that a second assessable dichotomous variable would have been relevant (i.e. presence or absence of vascular invasion), the addition of a second variable would have substantially increased the number of required observations and events in each additional subgroup. Therefore, while the sample size (n= 655) of our study was much larger than previously reported studies of T1 cN0 SCCP (Table 1) [8–11], it could have benefited from an even larger sample size.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Pierre I. Karakiewicz is partially supported by the University of Montreal Health Center Urology Specialists, Fonds de la Recherche en Santé du Quebec, the University of Montreal Department of Surgery and the University of Montreal Health Center (CHUM) Foundation. Rodolphe Thuret is partially supported by the Association Française d'Urologie (AFU).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES